ORIGINAL ARTICLE
Variation in Utilization of Health Care Services for Rural VA Enrollees With Mental Health-Related Diagnoses Christopher E. Johnson, PhD;1 Ruth L. Bush, MD, MPH;2 Jeffrey Harman, PhD;3 Jane Bolin, RN, JD, PhD;4 Gina Evans Hudnall, PhD;5,6 & Ann M. Nguyen, MPH1 1 Department of Health Services, School of Public Health, University of Washington, Seattle, Washington 2 College of Medicine, Texas A&M Health Science Center, Round Rock, Texas 3 Department of Health Services Research, Management, and Policy, College of Public Health and Health Professions, University of Florida, Gainesville, Florida 4 Southwest Rural Health Research Center, Texas A&M Health Science Center, College Station, Texas 5 South Central Mental Illness, Research, Education and Clinical Center and Houston Center for Quality of Care and Utilization Studies, Michael E. Debakey VA Medical Center, Houston, Texas 6 Health Services Research and Development Section, Department of Medicine, Baylor College of Medicine, Houston, Texas
Abstract Disclosures: The authors have no disclosures to report. Funding: The research reported here was partially supported by the U.S. Department of Veterans Affairs, Veterans Health Administration, VA Office of Rural Health (project IPA 0005471, Christopher Johnson, PI). The views expressed in this report are those of the authors and do not necessarily represent the views of the Department of Veterans Affairs. For further information, contact: Christopher E. Johnson, PhD, Department of Health Services, School of Public Health, University of Washington, H660D, Health Sciences Building, Box 357660, Seattle, WA 98195-7660; e-mail: [email protected]. doi: 10.1111/jrh.12105
Purpose: Rural-dwelling Department of Veterans Affairs (VA) enrollees are at high risk for a wide variety of mental health-related disorders. The objective of this study is to examine the variation in the types of mental and nonmental health services received by rural VA enrollees who have a mental healthrelated diagnosis. Methods: The Andersen and Aday behavioral model of health services use and the Agency for Healthcare Research and Quality Medical Expenditure Panel Survey (MEPS) data were used to examine how VA enrollees with mental health-related diagnoses accessed places of care from 1999 to 2009. Population survey weights were applied to the MEPS data, and logit regression was conducted to model how predisposing, enabling, and need factors influence rural veteran health services use (measured by visits to different places of care). Analyses were performed on the subpopulations: rural VA, rural nonVA, urban VA, and urban non-VA enrollees. Findings: For all types of care, both rural and urban VA enrollees received care from inpatient, outpatient, office-based, and emergency room settings at higher odds than urban non-VA enrollees. Rural VA enrollees also received all types of care from inpatient, office-based, and emergency room settings at higher odds than urban VA enrollees. Rural VA enrollees had higher odds of a mental health visit of any kind compared to urban VA and non-VA enrollees. Conclusions: Based on these variations, the VA may want to develop strategies to increase screening efforts in inpatient settings and emergency rooms to further capture rural VA enrollees who have undiagnosed mental health conditions. Key words access to care, health services research, mental health, utilization of health services.
The changing demographics of the US military and the increased reliance on the reserve and National Guard mean that more and more military men and women are being deployed into combat areas. During wartime
244
conflicts, a higher proportion of US military recruits tend to be from rural areas than from the general civilian population.1 As a result, there are higher proportions of veterans with combat experience residing in rural areas.
c 2015 National Rural Health Association The Journal of Rural Health 31 (2015) 244–253
Johnson et al.
Rural residents of all types face mental health care issues related to distance of treatment, provider shortages, treatment stigmas, and poverty.2-4 The rural-dwelling veteran population, which represents approximately 40% of the estimated 23,977,000 veterans in the United States, is at an even higher risk than nonveteran rural residents for a wide variety of mental health-related issues.5 These conditions often also impose stress upon rural veterans’ family members and caregivers. Unfortunately, access difficulties—such as those influenced by predisposing, enabling, and need factors—can impede the ability of these individuals to benefit from mental health care services.6 The implications of this from a public mental health care utilization perspective are cause for serious concern. Identifying the variation in types of health care services that veterans with mental health diagnoses receive could present opportunities for expanding access to assessments and services to rural veterans. Few empirical studies have examined the types of mental health and nonmental health services used by rural veterans with mental health diagnoses. Studies have examined the volume of nonmental health services care utilization and the impact of different settings on mental health services outcomes.7,8 None, however, have examined the access patterns for both types of services in the same population for all utilization related to the Department of Veterans Affairs (VA). This is an important issue since, of the over 8 million veterans enrolled in the Veterans Health Administration, approximately 36.4% (3,031,357) live in rural areas and roughly 1.5% (125,588) live in highly rural areas.9 Given the increased pressure to improve access to places of care, the VA implemented guidelines for improving mental health services for veterans through increased access to nonhospital-based settings.10,11 Recent studies of mental health services provided within the VA indicate that there is higher use of outpatient settings to provide mental health care to veterans and that distance to providers is negatively impacting long-term continuity of care.7 These results are further amplified by a study on the general rural population, which found that outpatient and other nontraditional settings are being used as gatekeepers to mental health services.12 In general, rural residents face a number of important mental health-related challenges compared to urban dwellers. They have higher levels of depression, substance abuse, domestic violence, incest, and child abuse.12,13 In rural regions, there is also low acceptability for mental health services due to increased stigma and decreased anonymity.14 They find it more difficult to locate and maintain health insurance coverage that provides for mental health conditions.4 A survey of 748,216 veterans found that rural dwellers within mental illness cohorts
c 2015 National Rural Health Association The Journal of Rural Health 31 (2015) 244–253
Health Services Utilization for Rural VA Enrollees
had worse physical and mental health component quality of life score summaries, and these differences were still significant after controlling for sociodemographic factors. The study also found that while rural veterans were less likely to have mental disorders, they had lower quality of life than urban veterans. These findings are consistent with other empirical analyses that show while rural dwellers in general have worse mental health services access and outcomes than the urban population, rural veterans are at an even greater disadvantage than their rural nonveteran counterparts.15,16 Gaps in mental health care are more likely for veterans who are younger, non-white, unmarried, homeless, and live in an area with limited access to VA facilities.17,18 Returning service personnel from the Iraq and Afghanistan theaters have documented problems with depression and posttraumatic stress disorder, and studies indicate the need for early intervention and treatment.19,20 The purpose of this study is to examine the variation in the types of mental and nonmental health services received by rural VA enrollees who have a mental health-related diagnosis, with particular interest in any differences between rural VA enrollees and rural non-VA enrollees. Based on evidence about utilization behavior in rural areas, the study hypothesizes that, when compared to urban non-VA adults with mental health-related expenditures, rural VA enrollees will have: (1) higher odds of accessing home health services, (2) lower odds of an inpatient stay, (3) higher odds of an outpatient physician visit, (4) higher odds of an office visit of any kind for health services, and (5) higher odds of an emergency room visit. These hypotheses were tested for both mental health services and all health services utilization. Doing this will shed light on what types of services veterans with mental health diagnoses receive differently from other populations.
Methods Data The Agency for Healthcare Research and Quality (AHRQ) Medical Expenditure Panel Survey (MEPS) public use data for 1999-2009 were used to examine longitudinal differences in patterns of utilization and sites of care for VA enrollees identified as having health care-related expenditures paid for by the VA. The data are collected annually in conjunction with the National Health Interview Survey. Both the MEPS Full-Year Consolidated Data Files and the Event Files capture data about the experiences of the population in accessing the health system. The Household Component (HC) of MEPS provides
245
Health Services Utilization for Rural VA Enrollees
a unique opportunity to analyze the services utilization of rural veterans (http://www.meps.ahrq.gov/mepsweb/). MEPS-HC is a longitudinal data set based on clustered and stratified samples of households that provide nationally representative estimates of health care use, insurance coverage, and sociodemographic characteristics for the US noninstitutionalized population.21 The data enable identification of VA and other types of insurance expenditures, the places where health care services are received, and the appropriate clinical code for each of these services. Places of care were identified by year through the MEPS Full-Year Consolidated Files that contain Household and Medical Provider Component data. Places of care included the following categories: home health visit, inpatient stay, outpatient physician visit, office-based visit, and emergency room visit. Table 1 contains the variable definitions. VA enrollees were identified in MEPS based on whether or not the VA paid for their health care services and if there was an expenditure source identified in the data. Survey respondents who had greater than zero total annual expenditures and were adults (18 or older) were included in these analyses. Rurality was determined using the MEPS metropolitan service area (MSA) variable. If the veteran lived outside of the MSA, he or she was categorized as living in a rural area. Whether or not the respondent had a mental health diagnosis was determined through the MEPS Medical Conditions Files and clinical classification codes (CCC). We included any person with a mental health condition, identified by CCC for adjustment disorders, anxiety disorders, attention deficit, delirium/dementia, impulse control disorders, mood disorders, schizophrenia, personality disorders, and alcohol-related disorders. Different coding was used for 2007-2009 MEPS (CCC 650-654 and 656-661), 2004-2006 MEPS (CCC 650-654 and 656-660), and 1999-2003 MEPS (CCC 066-072 and 074-075) due to changes in the CCC for those years. If a respondent saw a provider for one of these CCC during the year, that person was coded as having a mental health-related visit during that year. Provider visits were classified by places of care using the annual visit variables found in the MEPS Full-Year Consolidated Files. Self-reported general health and mental health status were summed over the 3 reporting periods during the year and averaged to represent a mean health status indicator for each respondent. These data have been used in previous studies designed to examine veterans’ utilization.22-24 After adjusting for adult subpopulation, mental health condition, health expenditures during the year, and missing values, there were 630
246
Johnson et al.
rural VA enrollees, 7,076 rural non-VA enrollees, 1,864 urban VA enrollees, and 27,456 urban non-VA enrollees (nonweighted observations) in the data analysis.
Conceptual Model To understand the variations in where rural VA enrollees with mental health-related diagnoses receive mental and nonmental health services, the Andersen and Aday behavioral model of health services utilization was used to form the conceptual model for the study.6 The conceptual model considered predisposing, enabling, and need variables, which are categories drawn from the Andersen and Aday model to define control variables for the rural VA enrollee population. The Andersen and Aday model has been used in other studies that examine health services utilization within the veteran population.25,26 Predisposing variables describe the social structure and biological influences on individual service utilization. These include health beliefs and sociodemographic characteristics. Regional and cultural influences on health care delivery have been documented in empirical studies; they have shown disparities in access to mental health services in rural regions where veterans travel an average of 45 miles from their residence to primary care clinics.27 Studies also show that cultural taboos are placed on obtaining mental health care within segments of society, especially among veterans regardless of geographic setting.12 Sociodemographic characteristics that influence how and where mental health care is utilized include: age, gender, race, education level, and marriage.26,28 In the data set, the following were used as predisposing variables for analysis: rural VA, rural non-VA, urban VA, urban non-VA (reference), age 18-34, age 35-64, age 65+ (reference), male, white, married, and education. The subpopulations and age groups were used as categorical variables. Male, white, and married were coded as dichotomous variables, while education was coded as continuous. Variable definitions are provided in Table 1. Enabling variables describe the resources available to individuals that would impact use of services. Access to income and employment with an employer offering health insurance coverage for mental conditions have been shown to increase health care options and utilization.4,29 The availability of any insurance coverage generally leads to higher health care utilization. In the data set, the following were used as enabling variables: total salary, employed, and any insurance. Total salary was coded as a continuous variable, while employed and any insurance were dichotomous. Need variables relate to the individual’s health status and how that will impact health services utilization.
c 2015 National Rural Health Association The Journal of Rural Health 31 (2015) 244–253
Johnson et al.
Health Services Utilization for Rural VA Enrollees
Table 1 Variable Definitions Variables Dependent variables Home health visit Inpatient stay Outpatient physician visit Office-based visit Emergency room visit Predisposing variables Rural VA Rural Non-VA Urban VA Urban Non-VA Age 18-34 Age 35-64 Age 65+ Male White Married Education Enabling variables Total salary Employed Any insurance Need variables Mental health status Health status Unable to get care Disability days
Definition
Whether or not the respondent had a home health visit during the year. Whether or not the respondent had an overnight stay in a hospital during the year. Whether or not the respondent had an outpatient physician visit during the year. Whether or not the respondent had an office-based visit of any kind during the year. Whether or not the respondent had an emergency room visit during the year. Whether or not the respondent’s care was paid for by the VA and other insurance and he or she lived outside of an MSA. Whether or not the respondent’s care was paid for by non-VA insurance and he or she lived outside of an MSA. Whether or not the respondent’s care was paid for by the VA and other insurance and he or she lived inside of an MSA. Whether or not the respondent’s care was paid for by non-VA and he or she lived inside of an MSA. Whether or not the respondent was between the ages of 18 and 34 during the year. Whether or not the respondent was between the ages of 35 and 64 during the year. Whether or not the respondent was age 65 or older during the year. Respondent was male. Respondent identified himself/herself as Caucasian. Respondent was married during the year. Highest degree/grade level respondent achieved. Total income earned by respondent during the year. Whether or not the respondent was employed any time during the year. Whether or not the respondent was covered by any health insurance. Mean self-reported mental health status for the year. Mean self-reported health status for the year. Whether or not the respondent could not access necessary care during the year. Respondent’s total number of disability days for the year.
Self-reported perceived health status and mental health status, reduced access to needed care, and number of annual sick leave or disability days are indicators of health care need.30 In the data set, the following were used as need variables: self-reported mental health status, self-reported health status, unable to get care, and disability days. These variables were coded as continuous, with the exception of unable to get care, which was dichotomous.
Data Analyses Survey weights were applied to each year included in the analyses to compute standard errors for MEPS-HC estimates.31 Survey weights were necessary due to the sampling design of MEPS-HC, which includes stratification, clustering, multiple stages of selection, and disproportionate sampling. Additionally, the sample weights reflect adjustments for survey nonresponse and adjustments to population control totals from the Current Population Survey. Stata 1232 was used to perform survey linearized weighted descriptive statistics for each subpopulation (rural VA, rural non-VA, urban
c 2015 National Rural Health Association The Journal of Rural Health 31 (2015) 244–253
VA, and urban non-VA enrollee), all of which contain only adults (18 and older) with medical expenditures and mental health diagnosis. Stata 12 was used to perform logit regressions on the subpopulations to model how predisposing, enabling, and need factors influence rural veteran health services use (measured by visits for any type of care and only mental health care) across the place of care (home health visit, inpatient stay, outpatient physician visit, office-based visit, and emergency room visit). Logit regression was chosen due to the dichotomous nature of the dependent variables (did the respondent have that type of visit during a given year).
Results Table 2 summarizes the descriptive statistics for the variables used for each of the 4 subpopulations. Emergency room visits were nonsignificant across the subpopulations. Urban non-VA enrollees were nonsignificant across all categories. Rural VA enrollees had more home health visits than any other subpopulation. Additionally, they were most likely to be married. Compared to the other subpopulations, rural VA enrollees reported the
247
Health Services Utilization for Rural VA Enrollees
Johnson et al.
Table 2 Descriptive Statistics (% or Mean) for Adults With Medical Expenditures and Mental Health Diagnoses by Place of Care Where They Received Mental Health Services and by Predisposing, Enabling, and Need Variables Used in the Models
Home health visit Inpatient stay Outpatient physician visit Office-based visit Emergency room visit Age 18-34 Age 35-64 Male White Married Education Total salary Employed Any insurance Mental health status Health status Unable to get care Disability days
Rural VA
Rural Non-VA
Urban VA
Urban Non-VA
4.26%∗∗ 3.25%∗∗∗ 5.86%∗∗∗ 42.9%∗∗ 2.45% 5.32%∗∗∗ 55.0%∗∗∗ 79.0%∗∗∗ 90.5%∗∗∗ 65.0%∗∗∗ 2.91∗∗∗ $23,005∗∗∗ 28.3%∗∗∗ 100%∗∗∗ 3.04∗∗∗ 3.28∗∗∗ 10.3% 0.86
3.01% 2.10% 1.72% 37.9%∗∗ 2.76% 20.9%∗∗∗ 58.2% 28.2% 91.8%∗∗∗ 51.9%∗∗∗ 2.99∗∗∗ $21,850∗∗∗ 42.4%∗∗∗ 89.4%∗∗∗ 2.77∗∗∗ 2.97∗∗∗ 14.4%∗∗∗ 1.49
3.73%∗∗∗ 3.28%∗∗∗ 6.71%∗∗∗ 43.1%∗∗ 2.21% 6.49%∗∗∗ 57.3%∗∗∗ 80.3%∗∗∗ 84.5%∗∗∗ 56.4%∗∗∗ 3.36∗ $29,591 29.6%∗∗∗ 100%∗∗∗ 2.89∗∗∗ 3.17∗∗∗ 12.2% 1.79∗∗∗
2.50% 1.83% 1.94% 42.4% 2.53% 24.8% 58.5% 30.6% 87.2% 46.2% 3.33 $29,711 49.6% 91.9% 2.67 2.76 13.6% 1.67
∗
P ࣘ .05. P ࣘ .01. ∗∗∗ P ࣘ .001. ∗∗
poorest perceived mental and general health status. Rural VA enrollees were least educated and most likely to be unemployed. Rural and urban VA enrollees were very similar across categories with the exception of race (higher proportions of rural VA enrollees were white), marriage (higher proportions of rural VA enrollees were married), salaries (urban VA enrollees had higher salaries), and ability to get health care (urban VA enrollees reported higher proportions being unable to get care). Rural non-VA enrollees were most likely to be white, reported the lowest salaries, were least likely to be insured, and were least likely to get care. Another set of descriptive statistics are presented in Table 3, showing the proportion of the adult population with at least 1 mental health-related visit during 19992009 at one of the places of care. Mental health visits remained fairly stable and comparable across the subpopulations for every year except 2005-2008, when there was a significant increase in the proportion of rural VA enrollees with at least 1 mental health-related visit to a place of care. Even though urban VA enrollees generally reported better mental health status than rural veterans, for some years there was no significant difference in the proportion of mental health-related visits between the 2 groups. For most years, there were higher proportions of rural VA enrollees with at least 1 mental health visit than urban VA enrollees, although these differences were only statistically significant for 2005-2009. Table 4 presents the results of logit regression analyses for places of care, considering all types of care (not just 248
Table 3 Proportion of All Adults With Medical Expenditures and At Least 1 Mental Health Visit During the Year Year
Rural VA
Rural Non-VA
Urban VA
Urban Non-VA
1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009
14.2% 16.1% 15.8% 18.9% 17.2% 17.6% 25.1%∗∗ 29.6%∗∗∗ 28.6%∗∗ 29.7%∗∗∗ 26.9%∗∗
14.2% 14.0% 15.4% 18.1% 19.6% 21.1% 21.2% 21.7% 20.9% 20.5% 20.7%
15.3% 15.5% 18.3%∗ 18.9% 19.0% 18.4% 20.9% 23.5%∗ 26.3%∗∗∗ 22.8% 23.2%
13.3% 14.6% 15.8% 17.8% 17.8% 20.2% 20.1% 20.2% 18.9% 19.4% 20.3%
∗
P ࣘ .05. P ࣘ .01. ∗∗∗ P ࣘ .001. ∗∗
mental health related). For the subpopulations, the odds ratios are given as compared to urban non-VA enrollees. Compared to urban non-VA enrollees, rural VA enrollees had about twice the odds of an office-based visit, 1.7 times the odds of an inpatient stay, and 1.7 times the odds of an outpatient physician visit, given that all other variables are the same. Although not quite as significant, rural VA enrollees had 37% higher odds of visiting an emergency room than urban non-VA enrollees. These trends hold for urban VA enrollees, though the exception is that an urban VA enrollee has only 1.5 times the odds of an inpatient stay. Overall, rural and urban VA enrollees presented similar patterns of utilization as c 2015 National Rural Health Association The Journal of Rural Health 31 (2015) 244–253
Johnson et al.
Health Services Utilization for Rural VA Enrollees
Table 4 Logit Regression Odds Ratios for Adults With Medical Expenditures and Mental Health Diagnoses and Places for all Types of Care, 1999-2009a
Rural VA Rural Non-VA Urban VA Urban Non-VA Age 18-34 Age 35-64 Age 65+ Male White Married Education Total salary Employed Any insurance Mental health status Health status Unable to get care Disability days
Home Health Visit
Inpatient Stay
Outpatient Physician Visit
Office-Based Visit
Emergency Room Visit
1.216 0.982 0.969 Ref. 0.160∗∗∗ 0.248∗∗∗ Ref. 0.935 0.766∗∗ 0.438∗∗∗ 1.004 0.999 0.303∗∗∗ 3.183∗∗∗ 1.085∗ 1.948∗∗∗ 0.952 1.026∗∗∗
1.731∗∗∗ 1.126∗ 1.469∗∗∗ Ref. 0.714∗∗∗ 0.580∗∗∗ Ref. 0.834∗∗∗ 0.991 0.984 1.011 0.999∗ 0.505∗∗∗ 1.758∗∗∗ 0.897∗∗∗ 1.802∗∗∗ 1.013 1.062∗∗∗
1.684∗∗∗ 1.066 1.847∗∗∗ Ref. 0.626∗∗∗ 0.958 Ref. 0.774∗∗∗ 0.977 1.111∗∗ 1.054∗∗∗ 1.000∗ 0.747∗∗∗ 2.266∗∗∗ 0.871∗∗∗ 1.454∗∗∗ 0.992 1.033∗∗∗
2.145∗∗ 0.983 1.901∗∗∗ Ref. 0.468∗∗∗ 0.588∗∗∗ Ref. 0.487∗∗∗ 1.316∗∗∗ 1.100 1.073∗∗∗ 1.000∗∗∗ 0.752∗∗∗ 2.721∗∗∗ 1.084∗ 1.322∗∗∗ 0.905 1.109∗∗∗
1.369∗∗ 1.127∗ 1.286∗∗∗ Ref. 1.295∗∗∗ 0.878∗∗ Ref. 0.860∗∗∗ 0.866∗∗∗ 0.798∗∗∗ 0.964∗∗∗ 0.999∗∗ 0.792∗∗∗ 1.091 0.932∗∗∗ 1.624∗∗∗ 1.234∗∗∗ 1.043∗∗∗
a
Year dummy variables for 2000-2009 not shown (1999 omitted for comparison). P ࣘ .05. ∗∗ P ࣘ .01. ∗∗∗ P ࣘ .001. ∗
urban non-VA enrollees. There was not as much similarity, however, when comparing subpopulations within the same geographic region. Rural VA enrollees had higher odds of health care usage than rural non-VA enrollees for office-based visits, inpatient stays, outpatient physician visits, and emergency room visits. Table 5 similarly presents the results of logit regression analyses for places of care; however, it only considers mental health-related services. Compared to urban non-VA enrollees, both rural and urban VA enrollees had higher odds of having a mental health-related visit of any type, given that all other variables are the same. The rural non-VA enrollee population had 8.2% lower odds of having any type of mental health visit and 20% lower odds of an office-based visit compared to urban non-VA enrollees. Rural VA enrollees had nearly 3 times the odds of seeing an outpatient physician compared to urban non-VA enrollees. Urban VA enrollees had 67% higher odds of an inpatient stay, 236% higher odds of an outpatient visit, and 20% higher odds of an office-based visit, compared to urban non-VA enrollees.
Discussion An unexpected result from this study is that rural VA enrollees appear to obtain care very differently from the general rural population. Rural VA enrollees had slightly higher odds of a mental health visit of any kind and an office-based visit for mental health services than rural c 2015 National Rural Health Association The Journal of Rural Health 31 (2015) 244–253
non-VA enrollees (Table 5). The rural VA population also had much higher odds for an outpatient visit for mental health services than the rural non-VA population (Table 5). When it came to visits to places for all types of care (not just mental health-related services), rural VA enrollees had higher odds of an inpatient stay, outpatient physician visit, office-based visit, and emergency room visit than rural non-VA enrollees (Table 4). A potential explanation for this difference that will require additional study is: Rural veterans may have better access to mental health and related services than the general rural population, due to the VA’s attempts to improve the availability of these services for this population. For example, the VA introduced the Veterans Crisis Line in 2007, a toll-free hotline for veterans and their family and friends, in response to rising suicide rates in the veteran population. The Veterans Crisis Line has since added online chat and text messaging. Additionally, the VA has been expanding its telemental health services, particularly in rural areas. In 2012, the VA initiated an aggressive Mental Health Hiring Initiative to add 1,600 mental health professionals and 300 support staff, a goal which was met in May 2013.33 Such improvements appear to be consistent with the significant increase in veterans’ use of mental health services beginning in 2005 shown in Table 3. While it is possible that mental health needs (not simply access) also differ between rural veterans and the general population, 1 study has suggested that need, in terms of suicide risk, is not significantly different between
249
Health Services Utilization for Rural VA Enrollees
Johnson et al.
Table 5 Logit Regression Odds Ratios for Adults With Medical Expenditures and Mental Health Diagnoses and Places of Care for Mental Health-Related Services, 1999-2009a
Rural VA Rural Non-VA Urban VA Urban Non-VA Age 18-34 Age 35-64 Age 65+ Male White Married Education Total salary Employed Any insurance Mental health status Health status Unable to get care Disability days
Any Type of Mental Health Visit
Home Health Visit
Inpatient Stay
Outpatient Visit
OfficeBased Visit
Emergency Visit
1.199∗∗ 0.918∗∗ 1.239∗∗∗ Ref. 1.457∗∗∗ 1.762∗∗∗ Ref. 0.544∗∗∗ 2.018∗∗∗ 0.685∗∗∗ 1.093∗∗∗ 1.000∗∗∗ 0.725∗∗∗ 1.550∗∗∗ 3.571∗∗∗ 0.880∗∗∗ 1.540∗∗∗ 1.016∗∗∗
0.964 1.016 0.973 Ref. 0.222∗∗∗ 0.343∗∗∗ Ref. 0.803∗ 0.956 0.257∗∗∗ 1.034 0.999 0.198∗∗∗ 3.176∗∗∗ 3.018∗∗∗ 1.297∗∗∗ 0.857 1.012
1.448 0.965 1.665∗∗ Ref. 2.275∗∗∗ 1.908∗∗∗ Ref. 1.025 1.077 0.526∗∗∗ 1.022 0.999∗∗ 0.451∗∗∗ 1.473∗∗ 4.283∗∗∗ 0.802∗∗∗ 1.478∗∗ 1.034∗∗∗
2.743∗∗∗ 0.826 3.364∗∗∗ Ref. 2.214∗∗∗ 2.685∗∗∗ Ref. 0.900 0.924 0.543∗∗∗ 1.096∗∗∗ 0.999 0.558∗∗∗ 2.496∗∗∗ 3.415∗∗∗ 0.790∗∗∗ 1.014 1.020∗∗∗
1.121 0.799∗∗∗ 1.199∗∗ Ref. 2.203∗∗∗ 2.365∗∗∗ Ref. 0.626∗∗∗ 2.016∗∗∗ 0.682∗∗∗ 1.095∗∗∗ 1.000∗∗∗ 0.665∗∗∗ 1.819∗∗∗ 4.452∗∗∗ 0.765∗∗∗ 1.292∗∗∗ 1.013∗∗∗
1.094 1.004 1.071 Ref. 2.876∗∗∗ 2.011∗∗∗ Ref. 0.785∗∗ 1.230∗ 0.627∗∗∗ 0.974 0.999∗ 0.752∗∗ 1.042 2.913∗∗∗ 0.919 1.497∗∗∗ 1.028∗∗∗
a
Year dummy variables for 2000-2009 not shown (1999 omitted for comparison). P ࣘ .05. ∗∗ P ࣘ .01. ∗∗∗ P ࣘ .001. ∗
the 2.34 Other access differences may be explained by further stratification, such as by deployment,35 age,36 and type of the mental health illness.37 This explanation, however, is beyond the scope of this study; yet, the additional emphasis on outpatient mental health services access does appear to be having a positive effect on veterans’ abilities to receive the health care they need. It should be noted that differences found between predisposing, enabling, and need variables (Table 2) were consistent with prior research,15,38 though no prior studies have examined these variables across these 4 subpopulations. Another plausible explanation for why rural VA enrollees receive care differently from the general rural population may be related to service connection status. Given that more veterans are now returning from deployment with combat experience and that many live in rural areas,2 there is a higher proportion of veterans in rural areas with service-connected disabilities who get service-connected benefits. Veterans with service-connected disabilities are 4 times more likely to use VA services than those without serviceconnected disabilities.39 This explanation is also beyond the scope of this study, as the MEPS data set does not provide information on a veteran’s service connection status.
250
There was no support for the hypotheses that rural VA enrollees with mental health conditions would have different mental health utilizations for home health, inpatient, office-based, or the emergency room when compared to urban non-VA enrollees (Table 5). Since relatively small amounts of home health services actually include a mental health component, it is not surprising that there was no significant difference between the use of this setting for mental health-related services when comparing rural VA and urban non-VA enrollees. For inpatient settings, there was also no significant difference in how rural VA enrollees and non-VA enrollees used this place of care for mental health-related services, despite there being potential distance-related access issues for rural VA enrollees.40 For office-based visits, however, it is surprising that rural VA enrollees do not use these clinics that are not attached to hospitals for mental health services differently than their urban non-VA enrollee counterparts. Given the difficulty some rural VA enrollees face in gaining access to hospital-based places of care, this subpopulation was expected to obtain treatment in office-based settings differently from the urban non-VA subpopulation. Office-based settings could be seen as substitutes for outpatient hospital treatment, especially in rural areas, but there does not appear to be a significant difference between their usages for mental
c 2015 National Rural Health Association The Journal of Rural Health 31 (2015) 244–253
Johnson et al.
health services when comparing rural VA and urban nonVA enrollees. Finally, emergency rooms were expected to be a last resort treatment source, but rural VA enrollees do not appear to use this setting for mental health services differently from the urban non-VA enrollees. The primary differences were found in overall mental health visits of any kind and outpatient physician visits for mental health services. First, it appears that rural VA enrollees use office-based visits for all types of care at slightly higher odds than urban VA enrollees and at 2 times higher odds than urban non-VA enrollees (Table 4). This may reflect an overall difference in demand for office-based visits and could potentially be the result of the VA’s heightened emphasis on integrating mental health guidelines into outpatient settings. Despite these higher odds, however, rural VA enrollees still visit hospitals to obtain mental health-related services differently from urban non-VA enrollees (Table 5). In fact, rural VA enrollees receive their mental health care during outpatient visits at almost 3 times higher odds than urban non-VA enrollees, even given the distance that rural VA enrollees need to travel to receive these services. This pattern holds when comparing urban VA to urban non-VA enrollees, with the former receiving their mental health care at over 3 times higher odds in the outpatient setting. It is also possible that differences in utilization between rural and urban VA enrollees are due to differences in perception of access.41 Aside from the outpatient setting, rural VA enrollees do not obtain mental health services in any other place of care at significantly higher odds than urban non-VA enrollees (Table 5). They do, however, use inpatient settings, office-based settings, and emergency rooms for all types of care at higher odds than the other subpopulations (Table 4). Inpatient settings and emergency rooms may be new areas where additional mental health screening could take place to further capture mental health conditions. This also shows that the higher use of inpatient and emergency rooms for all types of care is not simply a function of living in a rural area; there is a marked difference between rural VA enrollees and rural non-VA enrollees, which may be explained by differences in access as described earlier. The difference between rural VA enrollees and urban non-VA enrollees may be explained by cultural differences and insurance coverage.12,14 The VA should continue to focus on outpatient-related solutions for mental health access for rural veterans, but it may also want to develop strategies to navigate veterans to these services from other care intake sources (like inpatient units and emergency rooms). Procedures for referring veterans to appropriate mental health settings and guidelines for nonmental health providers could be developed to improve access for veterans who need mental health care but go either undiagnosed or untreated. c 2015 National Rural Health Association The Journal of Rural Health 31 (2015) 244–253
Health Services Utilization for Rural VA Enrollees
There were a few limitations associated with this study. The data used represent a national sample, but MEPS may under-represent VA utilization of places of care and under-sample veterans with mental health care diagnoses. Utilization reports are based on household samples, and there is some under-reporting of medical events. Respondents may have limited technical knowledge about the conditions they are reporting. An aggregate measure (the CCC) was used to capture the medical conditions for the respondents, and while it provided the level of categorization required for the study, full ICD-9 codes may provide a clearer picture; this level of detail was not reported in the MEPS public use files. Because of the exploratory nature of this analysis, the Andersen and Aday model was used for variable selection, but the bi-directional effects were not examined in this study. The use of MSA to define rural dwellers does not provide the optimal geographic identifier to distinguish veterans who live in highly rural areas versus those who are less rural. Unfortunately, ZIP codes that would allow for this higher level of detail are not available in the MEPS public use files. Finally, the supply of mental health services is difficult to measure. The American Medical Association has a database of psychiatrists; however, there is no comparable national database for other mental health professions.42 Despite these limitations, this study presents strong evidence about the utilization patterns for rural VA enrollees with mental health diagnoses. Specifically, the analyses point to: (1) rural VA enrollees receiving care differently from the general rural population; (2) rural and urban VA enrollees receiving mental health-related services from places of care in a similar fashion, with the exception of outpatient settings; and (3) places of care where rural VA enrollees receive all types of care being potentially different than where they primarily receive mental health services—this offers intervention opportunities for VA practitioners to increase access. Future analyses should examine interventions designed to identify and move newly diagnosed mental health patients from alternative access points to appropriate treatment. The VA may want to develop new strategies that increase screening efforts in inpatient and emergency rooms to further capture undiagnosed rural VA enrollees with mental health conditions. Future research should also examine use of mental health services by specific or grouped mental health diagnoses. Documented differences in mental health utilization by psychiatric diagnosis suggests that young adults with schizophrenia and posttraumatic stress disorder, and adults of all ages with bipolar disorder, have a high risk of not receiving general medical services. It would be interesting to see if such a difference exists in the receipt of mental health services.37 Finally, greater analyses of rurality may provide further strategies for 251
Health Services Utilization for Rural VA Enrollees
Johnson et al.
targeting interventions. Mental health has been reported to deteriorate as level of rurality increases,43,44 so veterans in highly rural areas may be in the greatest need of an intervention. 15.
References 1. Kane T. Who Bears the Burden?: Demographic Characteristics of US Military Recruits Before and After 9/11. Center for Data Analysis Report #05-08. Washington, DC: The Heritage; 2005. 2. Reschovsky JD, Staiti AB. Access and quality: does rural America lag behind? Health Aff (Millwood). 2005;24(4):1128-1139. 3. Ziller EC, Anderson NJ, Coburn AF. Access to rural mental health services: service use and out-of-pocket costs. J Rural Health. 2010;26(3):214-224. 4. Happell B, Scott D, Platania-Phung C, Nankivell J. Rural physical health care services for people with serious mental illness: a nursing perspective. Aust J Rural Health. 2012;20(5):248-253. 5. Rural Assistance Center. Returning Soldier and Veteran Health Frequently Asked Questions; 2013. Available at: http://www.raconline.org/topics/veterans/veteransfaq.php. Accessed December 6, 2013. 6. Andersen R. A Behavioral Model of Families’ Use of Health Services. Chicago, IL: University of Chicago, Center for Health Administration Studies; 1968. 7. King WR. Organizational transformation. Inform Syst Manage. 1997;14(2):63-65. 8. Jordan N, Lee TA, Valenstein M, Weiss KB. Effect of care setting on evidence-based depression treatment for veterans with COPD and comorbid depression. J Gen Intern Med. 2007;22(10):1447-1452. 9. West AN, Lee RE, Shambaugh-Miller MD, et al. Defining “rural” for veterans’ health care planning. J Rural Health. 2010;26(4):301-309. 10. Greenberg GA, Rosenheck RA. An evaluation of an initiative to improve Veterans Health Administration mental health services: broad impacts of the VHA’s Mental Health Strategic Plan. Mil Med. 2009;174(12):1263-1269. 11. U.S. Department of Veterans Affairs. Uniform Mental Health Services in VA Medical Centers and Clinics. Washington, DC: Veterans Health Administration; 2008. 12. Smalley KB, Yancey CT, Warren JC, Naufel K, Ryan R, Pugh JL. Rural mental health and psychological treatment: a review for practitioners. J Clin Psychol. 2010;66(5):479-489. 13. Bushy A. Health issues of women in rural environments: an overview. J Am Med Womens Assoc. 1997;53(2): 53-56. 14. Mohatt D, Adams S, Bradley M, Morris C. Mental Health and Rural America: 1994–2005. Rockville, MD: U.S.
252
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
Department of Health and Human Services, Health Resources and Services Administration, Office of Rural Health Policy; 2006. Available at: ftp://ftp.hrsa.gov/ruralhealth/RuralMentalHealth.pdf. Accessed December 6, 2013. Wallace AE, Weeks WB, Wang S, Lee AF, Kazis LE. Rural and urban disparities in health-related quality of life among veterans with psychiatric disorders. Psychiatr Serv. 2006;57(6):851-856. Mackenzie TA, Wallace AE, Weeks WB. Impact of rural residence on survival of male Veterans Affairs patients after age 65. J Rural Health. 2010;26(4):318-324. Blow FC, Zeber JE, McCarthy JF, Valenstein M, Gillon L, Bingham CR. Ethnicity and diagnostic patterns in veterans with psychoses. Soc Psychiatry Psychiatr Epidemiol. 2004;39(10):841-851. McCarthy JF, Valenstein M, Blow FC. Residential mobility among patients in the VA health system: associations with psychiatric morbidity, geographic accessibility, and continuity of care. Adm Policy Ment Health. 2007;34(5):448-455. Seal KH, Bertenthal D, Miner CR, Sen S, Marmar C. Bringing the war back home—mental health disorders among 103 788 US veterans returning from Iraq and Afghanistan seen at Department of Veterans Affairs facilities. Arch Intern Med. 2007;167(5):476-482. Eibner C. Invisible Wounds of War: Quantifying the Societal Costs of Psychological and Cognitive Injuries. Santa Monica, CA: RAND Corporation; 2008. Cohen JW, Cohen SB, Banthin JS. The medical expenditure panel survey: a national information resource to support healthcare cost research and inform policy and practice. Med Care. 2009;47(7 Suppl 1): S44-S50. West AN, Weeks WB. Who pays when VA users are hospitalized in the private sector?: evidence from three data sources. Med Care. 2007;45(10):1003-1007. Lasser KE, Woolhandler S, Himmelstein DU. Sources of US physician income: the contribution of government payments to the specialist–generalist income gap. J Gen Intern Med. 2008;23(9):1477-1481. Liu CF, Bolkan C, Chan D, Yano EM, Rubenstein LV, Chaney EF. Dual use of VA and non-VA services among primary care patients with depression. J Gen Intern Med. 2009;24(3):305-311. Elhai JD, Grubaugh AL, Richardson JD, Egede LE, Creamer M. Outpatient medical and mental healthcare utilization models among military veterans: results from the 2001 National Survey of Veterans. J Psychiatr Res. 2008;42(10):858-867. Fasoli DR, Glickman ME, Eisen SV. Predisposing characteristics, enabling resources and need as predictors of utilization and clinical outcomes for veterans receiving mental health services. Med Care. 2010;48(4):288295.
c 2015 National Rural Health Association The Journal of Rural Health 31 (2015) 244–253
Johnson et al.
27. Wallace AE, Lee R, Mackenzie TA, et al. A longitudinal analysis of rural and urban veterans’ health-related quality of life. J Rural Health. 2010;26(2):156-163. 28. Brennan PL, Kagay CR, Geppert JJ, Moos RH. Predictors and outcomes of outpatient mental health care—a 4-year prospective study of elderly medicare patients with substance use disorders. Med Care. 2001;39(1):39-49. 29. Sayer NA, Clothier B, Spoont M, Nelson DB. Use of mental health treatment among veterans filing claims for posttraumatic stress disorder. J Trauma Stress. 2007;20(1):15-25. 30. Comtois KA, Russo J, Snowden M, Srebnik D, Ries R, Roy-Byrne P. Factors associated with high use of public mental health services by persons with borderline personality disorder. Psychiatr Serv. 2003;54(8):1149-1154. 31. Machlin S, Yu W, Zodet M. Computing Standard Errors for MEPS Estimates. Rockville, MD: Agency for Healthcare Research and Quality; 2005. 32. Stata Reference Manual Release 7 [computer program]. Version. College Station, TX; 2001. 33. US Department of Veterans Affairs. Veterans and Mental Health; 2014. Available at: http://www.va.gov/opa/issues/mental_health.asp. Accessed July 10, 2014. 34. Kang H, Bullman T. Risk of suicide among US veterans after returning from the Iraq or Afghanistan war zones. J Am Med Assoc. 1990;300(6):652-653. 35. Hoge CW, Castro CA, Messer SC, McGurk D, Cotting DI, Koffman RL. Combat duty in Iraq and Afghanistan, mental health problems, and barriers to care. N Engl J Med. 2004;351(1):13-22. 36. Kazis LE, Miller DR, Clark J, et al. Health-related quality of life in patients served by the Department of Veterans
c 2015 National Rural Health Association The Journal of Rural Health 31 (2015) 244–253
Health Services Utilization for Rural VA Enrollees
37.
38.
39.
40.
41.
42.
43.
44.
Affairs: results from the Veterans Health Study. Arch Intern Med. 1998;158(6):626-632. Cradock-O’Leary J, Young AS, Yano EM, Wang MM, Lee ML. Use of general medical services by VA patients with psychiatric disorders. Psychiatr Serv. 2002;53(7):874878. Weeks WB, Kazis LE, Shen Y, et al. Differences in health-related quality of life in rural and urban veterans. Am J Public Health. 2004;94(10):1762-1767. US General Accounting Office. Report to the Ranking Minority Member, Subcommittee on Compensation, Pension, Insurance and Memorial Affairs, Committee on Veterans’ Affairs, House of Representatives: VA Health Care: How Distance from VA Facilities Affects Veterans’ Use of VA Services. Washington, DC: Health, Education, and Human Services Division; 1996. Weeks WB, Wallace AE, Wang S, Lee A, Kazis LE. Rural–urban disparities in health-related quality of life within disease categories of veterans. J Rural Health. 2006;22(3):204-211. Rost K, Fortney J, Zhang M, Smith J, Smith GR Jr. Treatment of depression in rural Arkansas: policy implications for improving care. J Rural Health. 1999;15(3):308-315. Thomas KC, Ellis AR, Konrad TR, Holzer CE, Morrissey JP. County-level estimates of mental health professional shortage in the United States. Psychiatr Serv. 2009;60(10):1323-1328. Eberhardt MS, Pamuk ER. The importance of place of residence: examining health in rural and nonrural areas. Am J Public Health. 2004;94(10):1682-1686. Hauenstein EJ, Petterson S, Merwin E, Rovnyak V, Heise B, Wagner D. Rurality, gender, and mental health treatment. Fam Community Health. 2006;29(3):169-185.
253
Variation in Utilization of Health Care Services for Rural VA Enrollees With Mental Health-Related Diagnoses Christopher E. Johnson, PhD;1 Ruth L. Bush, MD, MPH;2 Jeffrey Harman, PhD;3 Jane Bolin, RN, JD, PhD;4 Gina Evans Hudnall, PhD;5,6 & Ann M. Nguyen, MPH1 1 Department of Health Services, School of Public Health, University of Washington, Seattle, Washington 2 College of Medicine, Texas A&M Health Science Center, Round Rock, Texas 3 Department of Health Services Research, Management, and Policy, College of Public Health and Health Professions, University of Florida, Gainesville, Florida 4 Southwest Rural Health Research Center, Texas A&M Health Science Center, College Station, Texas 5 South Central Mental Illness, Research, Education and Clinical Center and Houston Center for Quality of Care and Utilization Studies, Michael E. Debakey VA Medical Center, Houston, Texas 6 Health Services Research and Development Section, Department of Medicine, Baylor College of Medicine, Houston, Texas
Abstract Disclosures: The authors have no disclosures to report. Funding: The research reported here was partially supported by the U.S. Department of Veterans Affairs, Veterans Health Administration, VA Office of Rural Health (project IPA 0005471, Christopher Johnson, PI). The views expressed in this report are those of the authors and do not necessarily represent the views of the Department of Veterans Affairs. For further information, contact: Christopher E. Johnson, PhD, Department of Health Services, School of Public Health, University of Washington, H660D, Health Sciences Building, Box 357660, Seattle, WA 98195-7660; e-mail: [email protected]. doi: 10.1111/jrh.12105
Purpose: Rural-dwelling Department of Veterans Affairs (VA) enrollees are at high risk for a wide variety of mental health-related disorders. The objective of this study is to examine the variation in the types of mental and nonmental health services received by rural VA enrollees who have a mental healthrelated diagnosis. Methods: The Andersen and Aday behavioral model of health services use and the Agency for Healthcare Research and Quality Medical Expenditure Panel Survey (MEPS) data were used to examine how VA enrollees with mental health-related diagnoses accessed places of care from 1999 to 2009. Population survey weights were applied to the MEPS data, and logit regression was conducted to model how predisposing, enabling, and need factors influence rural veteran health services use (measured by visits to different places of care). Analyses were performed on the subpopulations: rural VA, rural nonVA, urban VA, and urban non-VA enrollees. Findings: For all types of care, both rural and urban VA enrollees received care from inpatient, outpatient, office-based, and emergency room settings at higher odds than urban non-VA enrollees. Rural VA enrollees also received all types of care from inpatient, office-based, and emergency room settings at higher odds than urban VA enrollees. Rural VA enrollees had higher odds of a mental health visit of any kind compared to urban VA and non-VA enrollees. Conclusions: Based on these variations, the VA may want to develop strategies to increase screening efforts in inpatient settings and emergency rooms to further capture rural VA enrollees who have undiagnosed mental health conditions. Key words access to care, health services research, mental health, utilization of health services.
The changing demographics of the US military and the increased reliance on the reserve and National Guard mean that more and more military men and women are being deployed into combat areas. During wartime
244
conflicts, a higher proportion of US military recruits tend to be from rural areas than from the general civilian population.1 As a result, there are higher proportions of veterans with combat experience residing in rural areas.
c 2015 National Rural Health Association The Journal of Rural Health 31 (2015) 244–253
Johnson et al.
Rural residents of all types face mental health care issues related to distance of treatment, provider shortages, treatment stigmas, and poverty.2-4 The rural-dwelling veteran population, which represents approximately 40% of the estimated 23,977,000 veterans in the United States, is at an even higher risk than nonveteran rural residents for a wide variety of mental health-related issues.5 These conditions often also impose stress upon rural veterans’ family members and caregivers. Unfortunately, access difficulties—such as those influenced by predisposing, enabling, and need factors—can impede the ability of these individuals to benefit from mental health care services.6 The implications of this from a public mental health care utilization perspective are cause for serious concern. Identifying the variation in types of health care services that veterans with mental health diagnoses receive could present opportunities for expanding access to assessments and services to rural veterans. Few empirical studies have examined the types of mental health and nonmental health services used by rural veterans with mental health diagnoses. Studies have examined the volume of nonmental health services care utilization and the impact of different settings on mental health services outcomes.7,8 None, however, have examined the access patterns for both types of services in the same population for all utilization related to the Department of Veterans Affairs (VA). This is an important issue since, of the over 8 million veterans enrolled in the Veterans Health Administration, approximately 36.4% (3,031,357) live in rural areas and roughly 1.5% (125,588) live in highly rural areas.9 Given the increased pressure to improve access to places of care, the VA implemented guidelines for improving mental health services for veterans through increased access to nonhospital-based settings.10,11 Recent studies of mental health services provided within the VA indicate that there is higher use of outpatient settings to provide mental health care to veterans and that distance to providers is negatively impacting long-term continuity of care.7 These results are further amplified by a study on the general rural population, which found that outpatient and other nontraditional settings are being used as gatekeepers to mental health services.12 In general, rural residents face a number of important mental health-related challenges compared to urban dwellers. They have higher levels of depression, substance abuse, domestic violence, incest, and child abuse.12,13 In rural regions, there is also low acceptability for mental health services due to increased stigma and decreased anonymity.14 They find it more difficult to locate and maintain health insurance coverage that provides for mental health conditions.4 A survey of 748,216 veterans found that rural dwellers within mental illness cohorts
c 2015 National Rural Health Association The Journal of Rural Health 31 (2015) 244–253
Health Services Utilization for Rural VA Enrollees
had worse physical and mental health component quality of life score summaries, and these differences were still significant after controlling for sociodemographic factors. The study also found that while rural veterans were less likely to have mental disorders, they had lower quality of life than urban veterans. These findings are consistent with other empirical analyses that show while rural dwellers in general have worse mental health services access and outcomes than the urban population, rural veterans are at an even greater disadvantage than their rural nonveteran counterparts.15,16 Gaps in mental health care are more likely for veterans who are younger, non-white, unmarried, homeless, and live in an area with limited access to VA facilities.17,18 Returning service personnel from the Iraq and Afghanistan theaters have documented problems with depression and posttraumatic stress disorder, and studies indicate the need for early intervention and treatment.19,20 The purpose of this study is to examine the variation in the types of mental and nonmental health services received by rural VA enrollees who have a mental health-related diagnosis, with particular interest in any differences between rural VA enrollees and rural non-VA enrollees. Based on evidence about utilization behavior in rural areas, the study hypothesizes that, when compared to urban non-VA adults with mental health-related expenditures, rural VA enrollees will have: (1) higher odds of accessing home health services, (2) lower odds of an inpatient stay, (3) higher odds of an outpatient physician visit, (4) higher odds of an office visit of any kind for health services, and (5) higher odds of an emergency room visit. These hypotheses were tested for both mental health services and all health services utilization. Doing this will shed light on what types of services veterans with mental health diagnoses receive differently from other populations.
Methods Data The Agency for Healthcare Research and Quality (AHRQ) Medical Expenditure Panel Survey (MEPS) public use data for 1999-2009 were used to examine longitudinal differences in patterns of utilization and sites of care for VA enrollees identified as having health care-related expenditures paid for by the VA. The data are collected annually in conjunction with the National Health Interview Survey. Both the MEPS Full-Year Consolidated Data Files and the Event Files capture data about the experiences of the population in accessing the health system. The Household Component (HC) of MEPS provides
245
Health Services Utilization for Rural VA Enrollees
a unique opportunity to analyze the services utilization of rural veterans (http://www.meps.ahrq.gov/mepsweb/). MEPS-HC is a longitudinal data set based on clustered and stratified samples of households that provide nationally representative estimates of health care use, insurance coverage, and sociodemographic characteristics for the US noninstitutionalized population.21 The data enable identification of VA and other types of insurance expenditures, the places where health care services are received, and the appropriate clinical code for each of these services. Places of care were identified by year through the MEPS Full-Year Consolidated Files that contain Household and Medical Provider Component data. Places of care included the following categories: home health visit, inpatient stay, outpatient physician visit, office-based visit, and emergency room visit. Table 1 contains the variable definitions. VA enrollees were identified in MEPS based on whether or not the VA paid for their health care services and if there was an expenditure source identified in the data. Survey respondents who had greater than zero total annual expenditures and were adults (18 or older) were included in these analyses. Rurality was determined using the MEPS metropolitan service area (MSA) variable. If the veteran lived outside of the MSA, he or she was categorized as living in a rural area. Whether or not the respondent had a mental health diagnosis was determined through the MEPS Medical Conditions Files and clinical classification codes (CCC). We included any person with a mental health condition, identified by CCC for adjustment disorders, anxiety disorders, attention deficit, delirium/dementia, impulse control disorders, mood disorders, schizophrenia, personality disorders, and alcohol-related disorders. Different coding was used for 2007-2009 MEPS (CCC 650-654 and 656-661), 2004-2006 MEPS (CCC 650-654 and 656-660), and 1999-2003 MEPS (CCC 066-072 and 074-075) due to changes in the CCC for those years. If a respondent saw a provider for one of these CCC during the year, that person was coded as having a mental health-related visit during that year. Provider visits were classified by places of care using the annual visit variables found in the MEPS Full-Year Consolidated Files. Self-reported general health and mental health status were summed over the 3 reporting periods during the year and averaged to represent a mean health status indicator for each respondent. These data have been used in previous studies designed to examine veterans’ utilization.22-24 After adjusting for adult subpopulation, mental health condition, health expenditures during the year, and missing values, there were 630
246
Johnson et al.
rural VA enrollees, 7,076 rural non-VA enrollees, 1,864 urban VA enrollees, and 27,456 urban non-VA enrollees (nonweighted observations) in the data analysis.
Conceptual Model To understand the variations in where rural VA enrollees with mental health-related diagnoses receive mental and nonmental health services, the Andersen and Aday behavioral model of health services utilization was used to form the conceptual model for the study.6 The conceptual model considered predisposing, enabling, and need variables, which are categories drawn from the Andersen and Aday model to define control variables for the rural VA enrollee population. The Andersen and Aday model has been used in other studies that examine health services utilization within the veteran population.25,26 Predisposing variables describe the social structure and biological influences on individual service utilization. These include health beliefs and sociodemographic characteristics. Regional and cultural influences on health care delivery have been documented in empirical studies; they have shown disparities in access to mental health services in rural regions where veterans travel an average of 45 miles from their residence to primary care clinics.27 Studies also show that cultural taboos are placed on obtaining mental health care within segments of society, especially among veterans regardless of geographic setting.12 Sociodemographic characteristics that influence how and where mental health care is utilized include: age, gender, race, education level, and marriage.26,28 In the data set, the following were used as predisposing variables for analysis: rural VA, rural non-VA, urban VA, urban non-VA (reference), age 18-34, age 35-64, age 65+ (reference), male, white, married, and education. The subpopulations and age groups were used as categorical variables. Male, white, and married were coded as dichotomous variables, while education was coded as continuous. Variable definitions are provided in Table 1. Enabling variables describe the resources available to individuals that would impact use of services. Access to income and employment with an employer offering health insurance coverage for mental conditions have been shown to increase health care options and utilization.4,29 The availability of any insurance coverage generally leads to higher health care utilization. In the data set, the following were used as enabling variables: total salary, employed, and any insurance. Total salary was coded as a continuous variable, while employed and any insurance were dichotomous. Need variables relate to the individual’s health status and how that will impact health services utilization.
c 2015 National Rural Health Association The Journal of Rural Health 31 (2015) 244–253
Johnson et al.
Health Services Utilization for Rural VA Enrollees
Table 1 Variable Definitions Variables Dependent variables Home health visit Inpatient stay Outpatient physician visit Office-based visit Emergency room visit Predisposing variables Rural VA Rural Non-VA Urban VA Urban Non-VA Age 18-34 Age 35-64 Age 65+ Male White Married Education Enabling variables Total salary Employed Any insurance Need variables Mental health status Health status Unable to get care Disability days
Definition
Whether or not the respondent had a home health visit during the year. Whether or not the respondent had an overnight stay in a hospital during the year. Whether or not the respondent had an outpatient physician visit during the year. Whether or not the respondent had an office-based visit of any kind during the year. Whether or not the respondent had an emergency room visit during the year. Whether or not the respondent’s care was paid for by the VA and other insurance and he or she lived outside of an MSA. Whether or not the respondent’s care was paid for by non-VA insurance and he or she lived outside of an MSA. Whether or not the respondent’s care was paid for by the VA and other insurance and he or she lived inside of an MSA. Whether or not the respondent’s care was paid for by non-VA and he or she lived inside of an MSA. Whether or not the respondent was between the ages of 18 and 34 during the year. Whether or not the respondent was between the ages of 35 and 64 during the year. Whether or not the respondent was age 65 or older during the year. Respondent was male. Respondent identified himself/herself as Caucasian. Respondent was married during the year. Highest degree/grade level respondent achieved. Total income earned by respondent during the year. Whether or not the respondent was employed any time during the year. Whether or not the respondent was covered by any health insurance. Mean self-reported mental health status for the year. Mean self-reported health status for the year. Whether or not the respondent could not access necessary care during the year. Respondent’s total number of disability days for the year.
Self-reported perceived health status and mental health status, reduced access to needed care, and number of annual sick leave or disability days are indicators of health care need.30 In the data set, the following were used as need variables: self-reported mental health status, self-reported health status, unable to get care, and disability days. These variables were coded as continuous, with the exception of unable to get care, which was dichotomous.
Data Analyses Survey weights were applied to each year included in the analyses to compute standard errors for MEPS-HC estimates.31 Survey weights were necessary due to the sampling design of MEPS-HC, which includes stratification, clustering, multiple stages of selection, and disproportionate sampling. Additionally, the sample weights reflect adjustments for survey nonresponse and adjustments to population control totals from the Current Population Survey. Stata 1232 was used to perform survey linearized weighted descriptive statistics for each subpopulation (rural VA, rural non-VA, urban
c 2015 National Rural Health Association The Journal of Rural Health 31 (2015) 244–253
VA, and urban non-VA enrollee), all of which contain only adults (18 and older) with medical expenditures and mental health diagnosis. Stata 12 was used to perform logit regressions on the subpopulations to model how predisposing, enabling, and need factors influence rural veteran health services use (measured by visits for any type of care and only mental health care) across the place of care (home health visit, inpatient stay, outpatient physician visit, office-based visit, and emergency room visit). Logit regression was chosen due to the dichotomous nature of the dependent variables (did the respondent have that type of visit during a given year).
Results Table 2 summarizes the descriptive statistics for the variables used for each of the 4 subpopulations. Emergency room visits were nonsignificant across the subpopulations. Urban non-VA enrollees were nonsignificant across all categories. Rural VA enrollees had more home health visits than any other subpopulation. Additionally, they were most likely to be married. Compared to the other subpopulations, rural VA enrollees reported the
247
Health Services Utilization for Rural VA Enrollees
Johnson et al.
Table 2 Descriptive Statistics (% or Mean) for Adults With Medical Expenditures and Mental Health Diagnoses by Place of Care Where They Received Mental Health Services and by Predisposing, Enabling, and Need Variables Used in the Models
Home health visit Inpatient stay Outpatient physician visit Office-based visit Emergency room visit Age 18-34 Age 35-64 Male White Married Education Total salary Employed Any insurance Mental health status Health status Unable to get care Disability days
Rural VA
Rural Non-VA
Urban VA
Urban Non-VA
4.26%∗∗ 3.25%∗∗∗ 5.86%∗∗∗ 42.9%∗∗ 2.45% 5.32%∗∗∗ 55.0%∗∗∗ 79.0%∗∗∗ 90.5%∗∗∗ 65.0%∗∗∗ 2.91∗∗∗ $23,005∗∗∗ 28.3%∗∗∗ 100%∗∗∗ 3.04∗∗∗ 3.28∗∗∗ 10.3% 0.86
3.01% 2.10% 1.72% 37.9%∗∗ 2.76% 20.9%∗∗∗ 58.2% 28.2% 91.8%∗∗∗ 51.9%∗∗∗ 2.99∗∗∗ $21,850∗∗∗ 42.4%∗∗∗ 89.4%∗∗∗ 2.77∗∗∗ 2.97∗∗∗ 14.4%∗∗∗ 1.49
3.73%∗∗∗ 3.28%∗∗∗ 6.71%∗∗∗ 43.1%∗∗ 2.21% 6.49%∗∗∗ 57.3%∗∗∗ 80.3%∗∗∗ 84.5%∗∗∗ 56.4%∗∗∗ 3.36∗ $29,591 29.6%∗∗∗ 100%∗∗∗ 2.89∗∗∗ 3.17∗∗∗ 12.2% 1.79∗∗∗
2.50% 1.83% 1.94% 42.4% 2.53% 24.8% 58.5% 30.6% 87.2% 46.2% 3.33 $29,711 49.6% 91.9% 2.67 2.76 13.6% 1.67
∗
P ࣘ .05. P ࣘ .01. ∗∗∗ P ࣘ .001. ∗∗
poorest perceived mental and general health status. Rural VA enrollees were least educated and most likely to be unemployed. Rural and urban VA enrollees were very similar across categories with the exception of race (higher proportions of rural VA enrollees were white), marriage (higher proportions of rural VA enrollees were married), salaries (urban VA enrollees had higher salaries), and ability to get health care (urban VA enrollees reported higher proportions being unable to get care). Rural non-VA enrollees were most likely to be white, reported the lowest salaries, were least likely to be insured, and were least likely to get care. Another set of descriptive statistics are presented in Table 3, showing the proportion of the adult population with at least 1 mental health-related visit during 19992009 at one of the places of care. Mental health visits remained fairly stable and comparable across the subpopulations for every year except 2005-2008, when there was a significant increase in the proportion of rural VA enrollees with at least 1 mental health-related visit to a place of care. Even though urban VA enrollees generally reported better mental health status than rural veterans, for some years there was no significant difference in the proportion of mental health-related visits between the 2 groups. For most years, there were higher proportions of rural VA enrollees with at least 1 mental health visit than urban VA enrollees, although these differences were only statistically significant for 2005-2009. Table 4 presents the results of logit regression analyses for places of care, considering all types of care (not just 248
Table 3 Proportion of All Adults With Medical Expenditures and At Least 1 Mental Health Visit During the Year Year
Rural VA
Rural Non-VA
Urban VA
Urban Non-VA
1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009
14.2% 16.1% 15.8% 18.9% 17.2% 17.6% 25.1%∗∗ 29.6%∗∗∗ 28.6%∗∗ 29.7%∗∗∗ 26.9%∗∗
14.2% 14.0% 15.4% 18.1% 19.6% 21.1% 21.2% 21.7% 20.9% 20.5% 20.7%
15.3% 15.5% 18.3%∗ 18.9% 19.0% 18.4% 20.9% 23.5%∗ 26.3%∗∗∗ 22.8% 23.2%
13.3% 14.6% 15.8% 17.8% 17.8% 20.2% 20.1% 20.2% 18.9% 19.4% 20.3%
∗
P ࣘ .05. P ࣘ .01. ∗∗∗ P ࣘ .001. ∗∗
mental health related). For the subpopulations, the odds ratios are given as compared to urban non-VA enrollees. Compared to urban non-VA enrollees, rural VA enrollees had about twice the odds of an office-based visit, 1.7 times the odds of an inpatient stay, and 1.7 times the odds of an outpatient physician visit, given that all other variables are the same. Although not quite as significant, rural VA enrollees had 37% higher odds of visiting an emergency room than urban non-VA enrollees. These trends hold for urban VA enrollees, though the exception is that an urban VA enrollee has only 1.5 times the odds of an inpatient stay. Overall, rural and urban VA enrollees presented similar patterns of utilization as c 2015 National Rural Health Association The Journal of Rural Health 31 (2015) 244–253
Johnson et al.
Health Services Utilization for Rural VA Enrollees
Table 4 Logit Regression Odds Ratios for Adults With Medical Expenditures and Mental Health Diagnoses and Places for all Types of Care, 1999-2009a
Rural VA Rural Non-VA Urban VA Urban Non-VA Age 18-34 Age 35-64 Age 65+ Male White Married Education Total salary Employed Any insurance Mental health status Health status Unable to get care Disability days
Home Health Visit
Inpatient Stay
Outpatient Physician Visit
Office-Based Visit
Emergency Room Visit
1.216 0.982 0.969 Ref. 0.160∗∗∗ 0.248∗∗∗ Ref. 0.935 0.766∗∗ 0.438∗∗∗ 1.004 0.999 0.303∗∗∗ 3.183∗∗∗ 1.085∗ 1.948∗∗∗ 0.952 1.026∗∗∗
1.731∗∗∗ 1.126∗ 1.469∗∗∗ Ref. 0.714∗∗∗ 0.580∗∗∗ Ref. 0.834∗∗∗ 0.991 0.984 1.011 0.999∗ 0.505∗∗∗ 1.758∗∗∗ 0.897∗∗∗ 1.802∗∗∗ 1.013 1.062∗∗∗
1.684∗∗∗ 1.066 1.847∗∗∗ Ref. 0.626∗∗∗ 0.958 Ref. 0.774∗∗∗ 0.977 1.111∗∗ 1.054∗∗∗ 1.000∗ 0.747∗∗∗ 2.266∗∗∗ 0.871∗∗∗ 1.454∗∗∗ 0.992 1.033∗∗∗
2.145∗∗ 0.983 1.901∗∗∗ Ref. 0.468∗∗∗ 0.588∗∗∗ Ref. 0.487∗∗∗ 1.316∗∗∗ 1.100 1.073∗∗∗ 1.000∗∗∗ 0.752∗∗∗ 2.721∗∗∗ 1.084∗ 1.322∗∗∗ 0.905 1.109∗∗∗
1.369∗∗ 1.127∗ 1.286∗∗∗ Ref. 1.295∗∗∗ 0.878∗∗ Ref. 0.860∗∗∗ 0.866∗∗∗ 0.798∗∗∗ 0.964∗∗∗ 0.999∗∗ 0.792∗∗∗ 1.091 0.932∗∗∗ 1.624∗∗∗ 1.234∗∗∗ 1.043∗∗∗
a
Year dummy variables for 2000-2009 not shown (1999 omitted for comparison). P ࣘ .05. ∗∗ P ࣘ .01. ∗∗∗ P ࣘ .001. ∗
urban non-VA enrollees. There was not as much similarity, however, when comparing subpopulations within the same geographic region. Rural VA enrollees had higher odds of health care usage than rural non-VA enrollees for office-based visits, inpatient stays, outpatient physician visits, and emergency room visits. Table 5 similarly presents the results of logit regression analyses for places of care; however, it only considers mental health-related services. Compared to urban non-VA enrollees, both rural and urban VA enrollees had higher odds of having a mental health-related visit of any type, given that all other variables are the same. The rural non-VA enrollee population had 8.2% lower odds of having any type of mental health visit and 20% lower odds of an office-based visit compared to urban non-VA enrollees. Rural VA enrollees had nearly 3 times the odds of seeing an outpatient physician compared to urban non-VA enrollees. Urban VA enrollees had 67% higher odds of an inpatient stay, 236% higher odds of an outpatient visit, and 20% higher odds of an office-based visit, compared to urban non-VA enrollees.
Discussion An unexpected result from this study is that rural VA enrollees appear to obtain care very differently from the general rural population. Rural VA enrollees had slightly higher odds of a mental health visit of any kind and an office-based visit for mental health services than rural c 2015 National Rural Health Association The Journal of Rural Health 31 (2015) 244–253
non-VA enrollees (Table 5). The rural VA population also had much higher odds for an outpatient visit for mental health services than the rural non-VA population (Table 5). When it came to visits to places for all types of care (not just mental health-related services), rural VA enrollees had higher odds of an inpatient stay, outpatient physician visit, office-based visit, and emergency room visit than rural non-VA enrollees (Table 4). A potential explanation for this difference that will require additional study is: Rural veterans may have better access to mental health and related services than the general rural population, due to the VA’s attempts to improve the availability of these services for this population. For example, the VA introduced the Veterans Crisis Line in 2007, a toll-free hotline for veterans and their family and friends, in response to rising suicide rates in the veteran population. The Veterans Crisis Line has since added online chat and text messaging. Additionally, the VA has been expanding its telemental health services, particularly in rural areas. In 2012, the VA initiated an aggressive Mental Health Hiring Initiative to add 1,600 mental health professionals and 300 support staff, a goal which was met in May 2013.33 Such improvements appear to be consistent with the significant increase in veterans’ use of mental health services beginning in 2005 shown in Table 3. While it is possible that mental health needs (not simply access) also differ between rural veterans and the general population, 1 study has suggested that need, in terms of suicide risk, is not significantly different between
249
Health Services Utilization for Rural VA Enrollees
Johnson et al.
Table 5 Logit Regression Odds Ratios for Adults With Medical Expenditures and Mental Health Diagnoses and Places of Care for Mental Health-Related Services, 1999-2009a
Rural VA Rural Non-VA Urban VA Urban Non-VA Age 18-34 Age 35-64 Age 65+ Male White Married Education Total salary Employed Any insurance Mental health status Health status Unable to get care Disability days
Any Type of Mental Health Visit
Home Health Visit
Inpatient Stay
Outpatient Visit
OfficeBased Visit
Emergency Visit
1.199∗∗ 0.918∗∗ 1.239∗∗∗ Ref. 1.457∗∗∗ 1.762∗∗∗ Ref. 0.544∗∗∗ 2.018∗∗∗ 0.685∗∗∗ 1.093∗∗∗ 1.000∗∗∗ 0.725∗∗∗ 1.550∗∗∗ 3.571∗∗∗ 0.880∗∗∗ 1.540∗∗∗ 1.016∗∗∗
0.964 1.016 0.973 Ref. 0.222∗∗∗ 0.343∗∗∗ Ref. 0.803∗ 0.956 0.257∗∗∗ 1.034 0.999 0.198∗∗∗ 3.176∗∗∗ 3.018∗∗∗ 1.297∗∗∗ 0.857 1.012
1.448 0.965 1.665∗∗ Ref. 2.275∗∗∗ 1.908∗∗∗ Ref. 1.025 1.077 0.526∗∗∗ 1.022 0.999∗∗ 0.451∗∗∗ 1.473∗∗ 4.283∗∗∗ 0.802∗∗∗ 1.478∗∗ 1.034∗∗∗
2.743∗∗∗ 0.826 3.364∗∗∗ Ref. 2.214∗∗∗ 2.685∗∗∗ Ref. 0.900 0.924 0.543∗∗∗ 1.096∗∗∗ 0.999 0.558∗∗∗ 2.496∗∗∗ 3.415∗∗∗ 0.790∗∗∗ 1.014 1.020∗∗∗
1.121 0.799∗∗∗ 1.199∗∗ Ref. 2.203∗∗∗ 2.365∗∗∗ Ref. 0.626∗∗∗ 2.016∗∗∗ 0.682∗∗∗ 1.095∗∗∗ 1.000∗∗∗ 0.665∗∗∗ 1.819∗∗∗ 4.452∗∗∗ 0.765∗∗∗ 1.292∗∗∗ 1.013∗∗∗
1.094 1.004 1.071 Ref. 2.876∗∗∗ 2.011∗∗∗ Ref. 0.785∗∗ 1.230∗ 0.627∗∗∗ 0.974 0.999∗ 0.752∗∗ 1.042 2.913∗∗∗ 0.919 1.497∗∗∗ 1.028∗∗∗
a
Year dummy variables for 2000-2009 not shown (1999 omitted for comparison). P ࣘ .05. ∗∗ P ࣘ .01. ∗∗∗ P ࣘ .001. ∗
the 2.34 Other access differences may be explained by further stratification, such as by deployment,35 age,36 and type of the mental health illness.37 This explanation, however, is beyond the scope of this study; yet, the additional emphasis on outpatient mental health services access does appear to be having a positive effect on veterans’ abilities to receive the health care they need. It should be noted that differences found between predisposing, enabling, and need variables (Table 2) were consistent with prior research,15,38 though no prior studies have examined these variables across these 4 subpopulations. Another plausible explanation for why rural VA enrollees receive care differently from the general rural population may be related to service connection status. Given that more veterans are now returning from deployment with combat experience and that many live in rural areas,2 there is a higher proportion of veterans in rural areas with service-connected disabilities who get service-connected benefits. Veterans with service-connected disabilities are 4 times more likely to use VA services than those without serviceconnected disabilities.39 This explanation is also beyond the scope of this study, as the MEPS data set does not provide information on a veteran’s service connection status.
250
There was no support for the hypotheses that rural VA enrollees with mental health conditions would have different mental health utilizations for home health, inpatient, office-based, or the emergency room when compared to urban non-VA enrollees (Table 5). Since relatively small amounts of home health services actually include a mental health component, it is not surprising that there was no significant difference between the use of this setting for mental health-related services when comparing rural VA and urban non-VA enrollees. For inpatient settings, there was also no significant difference in how rural VA enrollees and non-VA enrollees used this place of care for mental health-related services, despite there being potential distance-related access issues for rural VA enrollees.40 For office-based visits, however, it is surprising that rural VA enrollees do not use these clinics that are not attached to hospitals for mental health services differently than their urban non-VA enrollee counterparts. Given the difficulty some rural VA enrollees face in gaining access to hospital-based places of care, this subpopulation was expected to obtain treatment in office-based settings differently from the urban non-VA subpopulation. Office-based settings could be seen as substitutes for outpatient hospital treatment, especially in rural areas, but there does not appear to be a significant difference between their usages for mental
c 2015 National Rural Health Association The Journal of Rural Health 31 (2015) 244–253
Johnson et al.
health services when comparing rural VA and urban nonVA enrollees. Finally, emergency rooms were expected to be a last resort treatment source, but rural VA enrollees do not appear to use this setting for mental health services differently from the urban non-VA enrollees. The primary differences were found in overall mental health visits of any kind and outpatient physician visits for mental health services. First, it appears that rural VA enrollees use office-based visits for all types of care at slightly higher odds than urban VA enrollees and at 2 times higher odds than urban non-VA enrollees (Table 4). This may reflect an overall difference in demand for office-based visits and could potentially be the result of the VA’s heightened emphasis on integrating mental health guidelines into outpatient settings. Despite these higher odds, however, rural VA enrollees still visit hospitals to obtain mental health-related services differently from urban non-VA enrollees (Table 5). In fact, rural VA enrollees receive their mental health care during outpatient visits at almost 3 times higher odds than urban non-VA enrollees, even given the distance that rural VA enrollees need to travel to receive these services. This pattern holds when comparing urban VA to urban non-VA enrollees, with the former receiving their mental health care at over 3 times higher odds in the outpatient setting. It is also possible that differences in utilization between rural and urban VA enrollees are due to differences in perception of access.41 Aside from the outpatient setting, rural VA enrollees do not obtain mental health services in any other place of care at significantly higher odds than urban non-VA enrollees (Table 5). They do, however, use inpatient settings, office-based settings, and emergency rooms for all types of care at higher odds than the other subpopulations (Table 4). Inpatient settings and emergency rooms may be new areas where additional mental health screening could take place to further capture mental health conditions. This also shows that the higher use of inpatient and emergency rooms for all types of care is not simply a function of living in a rural area; there is a marked difference between rural VA enrollees and rural non-VA enrollees, which may be explained by differences in access as described earlier. The difference between rural VA enrollees and urban non-VA enrollees may be explained by cultural differences and insurance coverage.12,14 The VA should continue to focus on outpatient-related solutions for mental health access for rural veterans, but it may also want to develop strategies to navigate veterans to these services from other care intake sources (like inpatient units and emergency rooms). Procedures for referring veterans to appropriate mental health settings and guidelines for nonmental health providers could be developed to improve access for veterans who need mental health care but go either undiagnosed or untreated. c 2015 National Rural Health Association The Journal of Rural Health 31 (2015) 244–253
Health Services Utilization for Rural VA Enrollees
There were a few limitations associated with this study. The data used represent a national sample, but MEPS may under-represent VA utilization of places of care and under-sample veterans with mental health care diagnoses. Utilization reports are based on household samples, and there is some under-reporting of medical events. Respondents may have limited technical knowledge about the conditions they are reporting. An aggregate measure (the CCC) was used to capture the medical conditions for the respondents, and while it provided the level of categorization required for the study, full ICD-9 codes may provide a clearer picture; this level of detail was not reported in the MEPS public use files. Because of the exploratory nature of this analysis, the Andersen and Aday model was used for variable selection, but the bi-directional effects were not examined in this study. The use of MSA to define rural dwellers does not provide the optimal geographic identifier to distinguish veterans who live in highly rural areas versus those who are less rural. Unfortunately, ZIP codes that would allow for this higher level of detail are not available in the MEPS public use files. Finally, the supply of mental health services is difficult to measure. The American Medical Association has a database of psychiatrists; however, there is no comparable national database for other mental health professions.42 Despite these limitations, this study presents strong evidence about the utilization patterns for rural VA enrollees with mental health diagnoses. Specifically, the analyses point to: (1) rural VA enrollees receiving care differently from the general rural population; (2) rural and urban VA enrollees receiving mental health-related services from places of care in a similar fashion, with the exception of outpatient settings; and (3) places of care where rural VA enrollees receive all types of care being potentially different than where they primarily receive mental health services—this offers intervention opportunities for VA practitioners to increase access. Future analyses should examine interventions designed to identify and move newly diagnosed mental health patients from alternative access points to appropriate treatment. The VA may want to develop new strategies that increase screening efforts in inpatient and emergency rooms to further capture undiagnosed rural VA enrollees with mental health conditions. Future research should also examine use of mental health services by specific or grouped mental health diagnoses. Documented differences in mental health utilization by psychiatric diagnosis suggests that young adults with schizophrenia and posttraumatic stress disorder, and adults of all ages with bipolar disorder, have a high risk of not receiving general medical services. It would be interesting to see if such a difference exists in the receipt of mental health services.37 Finally, greater analyses of rurality may provide further strategies for 251
Health Services Utilization for Rural VA Enrollees
Johnson et al.
targeting interventions. Mental health has been reported to deteriorate as level of rurality increases,43,44 so veterans in highly rural areas may be in the greatest need of an intervention. 15.
References 1. Kane T. Who Bears the Burden?: Demographic Characteristics of US Military Recruits Before and After 9/11. Center for Data Analysis Report #05-08. Washington, DC: The Heritage; 2005. 2. Reschovsky JD, Staiti AB. Access and quality: does rural America lag behind? Health Aff (Millwood). 2005;24(4):1128-1139. 3. Ziller EC, Anderson NJ, Coburn AF. Access to rural mental health services: service use and out-of-pocket costs. J Rural Health. 2010;26(3):214-224. 4. Happell B, Scott D, Platania-Phung C, Nankivell J. Rural physical health care services for people with serious mental illness: a nursing perspective. Aust J Rural Health. 2012;20(5):248-253. 5. Rural Assistance Center. Returning Soldier and Veteran Health Frequently Asked Questions; 2013. Available at: http://www.raconline.org/topics/veterans/veteransfaq.php. Accessed December 6, 2013. 6. Andersen R. A Behavioral Model of Families’ Use of Health Services. Chicago, IL: University of Chicago, Center for Health Administration Studies; 1968. 7. King WR. Organizational transformation. Inform Syst Manage. 1997;14(2):63-65. 8. Jordan N, Lee TA, Valenstein M, Weiss KB. Effect of care setting on evidence-based depression treatment for veterans with COPD and comorbid depression. J Gen Intern Med. 2007;22(10):1447-1452. 9. West AN, Lee RE, Shambaugh-Miller MD, et al. Defining “rural” for veterans’ health care planning. J Rural Health. 2010;26(4):301-309. 10. Greenberg GA, Rosenheck RA. An evaluation of an initiative to improve Veterans Health Administration mental health services: broad impacts of the VHA’s Mental Health Strategic Plan. Mil Med. 2009;174(12):1263-1269. 11. U.S. Department of Veterans Affairs. Uniform Mental Health Services in VA Medical Centers and Clinics. Washington, DC: Veterans Health Administration; 2008. 12. Smalley KB, Yancey CT, Warren JC, Naufel K, Ryan R, Pugh JL. Rural mental health and psychological treatment: a review for practitioners. J Clin Psychol. 2010;66(5):479-489. 13. Bushy A. Health issues of women in rural environments: an overview. J Am Med Womens Assoc. 1997;53(2): 53-56. 14. Mohatt D, Adams S, Bradley M, Morris C. Mental Health and Rural America: 1994–2005. Rockville, MD: U.S.
252
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
Department of Health and Human Services, Health Resources and Services Administration, Office of Rural Health Policy; 2006. Available at: ftp://ftp.hrsa.gov/ruralhealth/RuralMentalHealth.pdf. Accessed December 6, 2013. Wallace AE, Weeks WB, Wang S, Lee AF, Kazis LE. Rural and urban disparities in health-related quality of life among veterans with psychiatric disorders. Psychiatr Serv. 2006;57(6):851-856. Mackenzie TA, Wallace AE, Weeks WB. Impact of rural residence on survival of male Veterans Affairs patients after age 65. J Rural Health. 2010;26(4):318-324. Blow FC, Zeber JE, McCarthy JF, Valenstein M, Gillon L, Bingham CR. Ethnicity and diagnostic patterns in veterans with psychoses. Soc Psychiatry Psychiatr Epidemiol. 2004;39(10):841-851. McCarthy JF, Valenstein M, Blow FC. Residential mobility among patients in the VA health system: associations with psychiatric morbidity, geographic accessibility, and continuity of care. Adm Policy Ment Health. 2007;34(5):448-455. Seal KH, Bertenthal D, Miner CR, Sen S, Marmar C. Bringing the war back home—mental health disorders among 103 788 US veterans returning from Iraq and Afghanistan seen at Department of Veterans Affairs facilities. Arch Intern Med. 2007;167(5):476-482. Eibner C. Invisible Wounds of War: Quantifying the Societal Costs of Psychological and Cognitive Injuries. Santa Monica, CA: RAND Corporation; 2008. Cohen JW, Cohen SB, Banthin JS. The medical expenditure panel survey: a national information resource to support healthcare cost research and inform policy and practice. Med Care. 2009;47(7 Suppl 1): S44-S50. West AN, Weeks WB. Who pays when VA users are hospitalized in the private sector?: evidence from three data sources. Med Care. 2007;45(10):1003-1007. Lasser KE, Woolhandler S, Himmelstein DU. Sources of US physician income: the contribution of government payments to the specialist–generalist income gap. J Gen Intern Med. 2008;23(9):1477-1481. Liu CF, Bolkan C, Chan D, Yano EM, Rubenstein LV, Chaney EF. Dual use of VA and non-VA services among primary care patients with depression. J Gen Intern Med. 2009;24(3):305-311. Elhai JD, Grubaugh AL, Richardson JD, Egede LE, Creamer M. Outpatient medical and mental healthcare utilization models among military veterans: results from the 2001 National Survey of Veterans. J Psychiatr Res. 2008;42(10):858-867. Fasoli DR, Glickman ME, Eisen SV. Predisposing characteristics, enabling resources and need as predictors of utilization and clinical outcomes for veterans receiving mental health services. Med Care. 2010;48(4):288295.
c 2015 National Rural Health Association The Journal of Rural Health 31 (2015) 244–253
Johnson et al.
27. Wallace AE, Lee R, Mackenzie TA, et al. A longitudinal analysis of rural and urban veterans’ health-related quality of life. J Rural Health. 2010;26(2):156-163. 28. Brennan PL, Kagay CR, Geppert JJ, Moos RH. Predictors and outcomes of outpatient mental health care—a 4-year prospective study of elderly medicare patients with substance use disorders. Med Care. 2001;39(1):39-49. 29. Sayer NA, Clothier B, Spoont M, Nelson DB. Use of mental health treatment among veterans filing claims for posttraumatic stress disorder. J Trauma Stress. 2007;20(1):15-25. 30. Comtois KA, Russo J, Snowden M, Srebnik D, Ries R, Roy-Byrne P. Factors associated with high use of public mental health services by persons with borderline personality disorder. Psychiatr Serv. 2003;54(8):1149-1154. 31. Machlin S, Yu W, Zodet M. Computing Standard Errors for MEPS Estimates. Rockville, MD: Agency for Healthcare Research and Quality; 2005. 32. Stata Reference Manual Release 7 [computer program]. Version. College Station, TX; 2001. 33. US Department of Veterans Affairs. Veterans and Mental Health; 2014. Available at: http://www.va.gov/opa/issues/mental_health.asp. Accessed July 10, 2014. 34. Kang H, Bullman T. Risk of suicide among US veterans after returning from the Iraq or Afghanistan war zones. J Am Med Assoc. 1990;300(6):652-653. 35. Hoge CW, Castro CA, Messer SC, McGurk D, Cotting DI, Koffman RL. Combat duty in Iraq and Afghanistan, mental health problems, and barriers to care. N Engl J Med. 2004;351(1):13-22. 36. Kazis LE, Miller DR, Clark J, et al. Health-related quality of life in patients served by the Department of Veterans
c 2015 National Rural Health Association The Journal of Rural Health 31 (2015) 244–253
Health Services Utilization for Rural VA Enrollees
37.
38.
39.
40.
41.
42.
43.
44.
Affairs: results from the Veterans Health Study. Arch Intern Med. 1998;158(6):626-632. Cradock-O’Leary J, Young AS, Yano EM, Wang MM, Lee ML. Use of general medical services by VA patients with psychiatric disorders. Psychiatr Serv. 2002;53(7):874878. Weeks WB, Kazis LE, Shen Y, et al. Differences in health-related quality of life in rural and urban veterans. Am J Public Health. 2004;94(10):1762-1767. US General Accounting Office. Report to the Ranking Minority Member, Subcommittee on Compensation, Pension, Insurance and Memorial Affairs, Committee on Veterans’ Affairs, House of Representatives: VA Health Care: How Distance from VA Facilities Affects Veterans’ Use of VA Services. Washington, DC: Health, Education, and Human Services Division; 1996. Weeks WB, Wallace AE, Wang S, Lee A, Kazis LE. Rural–urban disparities in health-related quality of life within disease categories of veterans. J Rural Health. 2006;22(3):204-211. Rost K, Fortney J, Zhang M, Smith J, Smith GR Jr. Treatment of depression in rural Arkansas: policy implications for improving care. J Rural Health. 1999;15(3):308-315. Thomas KC, Ellis AR, Konrad TR, Holzer CE, Morrissey JP. County-level estimates of mental health professional shortage in the United States. Psychiatr Serv. 2009;60(10):1323-1328. Eberhardt MS, Pamuk ER. The importance of place of residence: examining health in rural and nonrural areas. Am J Public Health. 2004;94(10):1682-1686. Hauenstein EJ, Petterson S, Merwin E, Rovnyak V, Heise B, Wagner D. Rurality, gender, and mental health treatment. Fam Community Health. 2006;29(3):169-185.
253
